containing 2.

5 3 105 PBMCs was added to each well the presence of 106 irradiated PBMCs from at least
after counting in the presence of trypan blue. The
plates were incubated at 37°C for 4 hours before
harvesting 20 ml of supernatant. Percentage lysis
was estimated as [(experimental counts 2 media
control) 3 100]/(detergent counts 2 media control).
Lysis of non-peptide controls was subtracted to give
peptide-specific lysis. Media controls were between
10 to 15% of detergent controls.
22. One tetramer-positive CD81 cell was cloned directly
into each well of a 96-well flat-bottomed plate (Nunc) in
26. B. Autran et al., Science 277, 112 (1997 ).
three donors, together with phytohemagglutinin (5 mg/
ml, Wellcome) in RPMI 1640 (Gibco) supplemented with
glutamine, penicillin, streptomycin, and 5% human se-
rum. After 4 days, Lymphocult-T (Biotest) was added to
a final concentration of 10%. After 3 weeks the cultures
were expanded into 24-well plates.
23. P. J. R. Goulder et al., J. Exp Med. 185, 1423 (1997 ).
24. A. Carmichael et al., ibid. 177, 249 (1993).
25. D. D. Ho et al., Nature 373, 123 (1995); X. Wei et al.,
ibid., p. 117.
27. N. Steven et al., J. Exp. Med. 185, 1605 (1997 ).
28. The assistance of the nursing staff at Rockefeller Hos-
pital and of J. Jin and J. Song in processing PBMC
samples is gratefully acknowledged. Supported by
grants from NIH (U01AI41534) and General Clinical Re-
search Center (GCRC) (MO1-RR00102). G.S.O.,
P.R.D., V.C., S.L.R.-J, and A.J.M. are funded by the
Medical Research Council (UK). S.B. and N.A.M. are
supported by the Wellcome Trust.
25 November 1997; accepted 29 January 1998

Biodiversity Assessment and scarabs) or, where the majority of the fauna
has not been adequately cataloged, a well-
Conservation Strategies known monophyletic unit (antlions). In one
instance (termites), only an incomplete set
Albert S. van Jaarsveld,* Stefanie Freitag, Steven L. Chown, of published data from a systematic survey
Caron Muller, Stephanie Koch, Heath Hull, Chuck Bellamy, was available, resulting in poor species cov-
Martin Krüger, Sebastian Endrödy-Younga, Mervyn W. Mansell, erage (19). In none of these cases was there
reason to presume that the species chosen
Clarke H. Scholtz are a nonrandom subset of the taxon as a
whole with regard to geographic distribution.
The efficient representation of all species in conservation planning is problematic. Often, Data from the Transvaal region (now in-
species distribution is assessed by dividing the land into a grid; complementary sets of cluding Gauteng, Mpumalanga, Northern,
grids, in which each taxon is represented at least once, are then sought. To determine and part of North-West provinces; South
if this approach provides useful surrogate information, species and higher taxon data for Africa) were mapped on a 25 km by 25 km
South African plants and animals were analyzed. Complementary species sets did not grid (n 5 474), and complementary sets for
coincide and overlapped little with higher taxon sets. Survey extent and taxonomic each of the taxa were identified by means of
knowledge did not affect this overlap. Thus, the assumptions of surrogacy, on which so a rarity-based algorithm (12). The study area
much conservation planning is based, are not supported. is about the size of the United Kingdom and
comprises 20% of the surface area of one of
the most species-rich countries in the world.
Richness hotspots and coldspots reflect the
Practical conservation uses surrogate infor- rarity, and higher taxon richness as biodi- top 5% of species-rich and species-poor 25-
mation, such as richness of indicator taxa, versity surrogate measures (“traditional” km squares, respectively (14). Rare species
endemism (taxa restricted to a given area), surrogates) has been explored, and the con- are defined as those occurring in less than 24
or higher taxon richness (that is, genus or sensus is that richness “hotspots” (highly squares (5% of 474 squares), and this rarity
family richness) to identify possible conser- species-rich areas) and “coldspots” (areas may be the consequence of a restricted range
vation areas (1–8). Although not universally poor in species) rarely coincide; nor do or inadequate sampling (20). The degree of
accepted (9), there is broad agreement that hotspots and rare (restricted range) taxa spatial overlap among complementary sets,
conservation areas should strive to sample generally coincide (6, 14–17). However, species richness (hotspots and coldspots),
regional features, a goal that is most effi- the surrogacy value of complementary sets and areas containing rare taxa is expressed by
ciently accomplished with complementary has not been assessed. Here, the relation the Jaccard coefficient (Table 1).
sets (10, 11). These are sets of grids that between traditional surrogate measures and As in previous studies (14), we found
contain all species in a taxon at least once complementary sets, as well as the degree of little overlap within taxa using measures of
(10, 12); the complementarity principle en- overlap among complementary sets across richness (hotspots and coldspots) and rarity
sures that conservation areas represent all taxa, is investigated. (21) (Fig. 1 and Table 1). The single ex-
species efficiently and that rare species are The study incorporated 9119 species, in- ception was richness hotspots and rarity
included (10). Although the outcome of cluding well-studied taxa that are frequently where the mean overlap was 50% (Table 1).
such a complementarity analysis provides a used as biodiversity indicators (4), such as This high value was due mostly to high
sound basis for the efficient conservation of vascular plants (Plantae), mammals (Mam- overlap values in plants and in phytopha-
the focal taxon, it is commonly assumed that malia), birds (Aves), and butterflies (Hespe- gous insects (plants, buprestids, and butter-
the outcome is more widely applicable to rioidea and Papilionoidea), and less well- flies all had overlap values exceeding 75%)
other taxa (13). known taxa, such as termites (Isoptera), ant- (Table 1). Speciose plant regions in south-
The value of species richness, species lions (Myrmeleontidae), buprestid beetles ern Africa include large numbers of rare
(Buprestidae), and scarabaeoid beetles (Scar- plant species (22), and patterns in plant
A. S. van Jaarsveld, S. Freitag, S. L. Chown, C. Muller, S. abaeoidea) (18). These taxa vary consider- diversity are often a good predictor of pat-
Kock, H. Hull, C. H. Scholtz, Department of Zoology and ably with regard to survey extent and taxo- terns in insect diversity (23). This may
Entomology, University of Pretoria, Pretoria 0002, South nomic knowledge. For example, birds are account, at least to some extent, for the
Africa.
C. Bellamy, M. Krüger, S. Endrödy-Younga, Transvaal surveyed in all grid cells and all species are high overlap values of richness hotspots and
Museum, Post Office Box 413, Pretoria 0001, South included, whereas ;20% of antlion species rarity observed within each of these taxa.
Africa. are included and these are surveyed in 8% of Overlap among taxa for richness hotspots
M. W. Mansell, Plant Protection Research Institute, Agri-
cultural Research Council, Private Bag X134, Pretoria the grid cells in the study area. Species that and coldspots is, respectively, highest be-
0001, South Africa. were chosen for inclusion in the poorly sur- tween butterflies and plants (24%), and scar-
* To whom correspondence should be addressed. E-mail: veyed taxa represent either the known fauna ab and buprestid beetles (13%) (24). Over-
albert@scientia.up.ac.za. for the region (for example, buprestids and lap among areas containing rare taxa is most

2106 SCIENCE z VOL. 279 z 27 MARCH 1998 z www.sciencemag.org

 REPORTS
Table 1. Percentage overlap among types of priority conservation areas, the Jaccard coefficient [number of grids shared/(number of additional grids
species-based complementary sets for different taxa, and complementary selected for taxon A 1 number of additional grids selected for taxon B)] 3
sets representing different taxonomic levels. The overlap was calculated with 100.

Scarab Buprestid
Comparisons/taxa Mammals Birds Plants Butterflies Termites Antlions
beetles beetles

Priority conservation areas

Richness hotspots versus rare species 29.2 18.0 82.6 77.8 23.8 60.0 6.7 80.0
Richness coldspots versus rare species 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Complementary sets versus richness hotspots 8.1 20.0 8.6 16.3 11.1 16.7 32.1 19.7
Complementary sets versus richness coldspots 0.0 0.0 1.0 2.9 0.0 9.7 0.0 2.0
Complementary sets versus rare species 21.4 30.0 8.2 16.3 11.1 16.7 10.7 12.7
Complementary species sets
Mammals 11.9 6.6 8.5 0.0 3.5 9.3 11.8
Birds 7.3 9.8 0.0 6.5 13.6 8.6
Plants 12.7 0.4 2.2 7.3 19.5
Butterflies 0.0 2.0 11.7 20.7
Termites 0.0 0.0 1.5
Antlions 10.0 2.9
Scarab beetles 14.3
Buprestid beetles
Complementary sets representing different taxonomic levels
Species versus genus 17.9 34.5 37.6 17.8 20.0 0.0 24.0 34.4
Species versus family 8.0 3.7 7.4 2.3 40.0 12.5 4.0 1.6

pronounced in mammals and birds (37%).
Nonetheless, all of these overlap values are
low, indicating that different taxa are speci-
ose, species-poor, or have their rare species
represented, in different grid cells (24).
The mean coincidence between comple-
mentary species sets and grids selected on the
basis of richness (hotspots and coldspots),
and between complementary sets and grids
containing rare taxa, is well below 20% (Fig.
1 and Table 1). The highest overlap in com-
plementary sets and richness hotspots is for
scarab beetles (32%) and birds (20%); this
overlap reached only 8% in mammals. Co-
incidence between complementary sets and
rare taxa was highest in mammals (30%).
Thus, grids selected for a single representa-
tion of each species tend not to be those with
excessively high or unusually low species
richness, nor do they include a dispropor-
tionate number of rare species (Table 1).
Pairwise comparisons of complementary
species sets reveal a mean overlap of less
than 10% (Fig. 2 and Table 1); maximum
overlap (21%) is between butterflies and
buprestid beetles. In multiple comparisons
of complementary sets, no grid cell was
shared by all taxa, and a maximum of six
taxa shared complementary grids (coinci-
dentally, n 5 6 grids shared). This further
emphasizes the lack of overlap of comple-
mentary sets across taxa. Thus, different
conservation areas are required to conserve
different taxa.
Complementary sets that represent gen-
era and families show little overlap with
species-based complementary sets across taxa
(,30%) (Fig. 2 and Table 1). Maximum
overlap between genus- and species-based
sets is for plants (38%) and birds (35%), taxa
that are well surveyed and systematically
well known (25), and for buprestid beetles
(34%), a group that has not been well sur-
veyed and in which many species remain
undescribed (18, 26). Similarly, overlap be-
tween family- and species-based sets is high-
est for termites (40%) and antlions (13%),
which are either poorly surveyed or repre-
sented by few species in this analysis. In
contrast, the overlap between well-surveyed
and taxonomically well-represented groups,
namely plants, birds, and mammals, was
minimal, at 7, 4, and 8% respectively (Table
1). Patterns of overlap based on complemen-
tary sets were also inconsistent between taxa
with changing hierarchical levels (for plants,
overlap declines from 38 to 7% from genus
to family level, whereas for termites there
was an increase from 20 to 40%). Thus,
Fig. 1. The degree of spatial overlap (mean 6 SD
of Jaccard coefficient) between conservation ar-
eas generated by means of different prioritization
criteria (species-based complementary areas,
richness hotspots and coldspots, areas contain-
ing rare taxa).
selecting conservation areas by genus- or
family-level data cannot result in efficient
species-level conservation.
Our results provide little support for
the notion that species complementary
sets are congruent across taxa or that com-
plementary sets are congruent with rich-
ness (hotspots, coldspots, or both) or areas
harboring rare taxa, or both. In addition,
our results suggest that the use of higher
taxa as surrogates (27) for species-based
complementary set selection holds little
promise at a scale relevant to practical
conservation planning. This largely under-
mines hopes for using “indicator taxa” or
higher taxon surrogate information as
biodiversity planning tools. These data
Fig. 2. The degree of spatial overlap (mean 6 SD
of Jaccard coefficient) among species-based
complementary sets across higher taxonomic
groupings (that is, species-based surrogacy) and
overlap between the species-based priority con-
servation sets and sets generated by means of
genus and family level data (that is, higher taxon
surrogacy).

www.sciencemag.org z SCIENCE z VOL. 279 z 27 MARCH 1998 2107

also support findings from a recent study
that adopted a different approach and was
conducted at a very different scale, yet
also concluded that the prospects for in-
dicator taxa are poor (28). Furthermore,
conservation areas identified by means of
traditional prioritization criteria [richness
hotspots and coldspots and areas contain-
ing rare taxa (21)] are unlikely to be useful
surrogates for representative complemen-
tary conservation networks. This lack of
coincidence between taxa, hierarchical lev-
els, and traditional criteria for priority con-
servation areas implies that all available spe-
cies-based information should be incorporat-
ed into regional conservation assessments
(6). Moreover, these results underscore the
value of sound species-related distribution
data for conservation planning and empha-
size the necessity for survey research in con-
servation biology (29).
REFERENCES AND NOTES
___________________________
1. P. B. Landres, J. Verner, J. W. Thomas, Conserv.
Biol. 2, 316 (1988).
2. R. F. Noss, ibid. 4, 355 (1990).
3. J. Curnutt, J. Lockwood, H.-K. Luh, P. Nott, G. Rus-
sel, Nature 367, 326 (1994).
4. D. L. Pearson, Philos. Trans. R. Soc. London Ser. B
345, 75 (1994).
5. P. H. Williams and C. J. Humphries, in Biodiversity,
K. J. Gaston, Ed. (Blackwell, Oxford, 1996), pp. 54 –
76.
iiii
6. K. J. Gaston, Progr. Phys. Geogr. 20, 105 (1996).
26. C. H. Scholtz and S. L. Chown, S. Afr. J. Sci. 91, 124 30. Supported by the Foundation for Research Devel-
(1995). opment, the University of Pretoria, the Transvaal
Museum, and the Agricultural Research Council.
27. P. H. Williams and K. J. Gaston, Biol. Conserv. 67,
211 (1994). K. J. Gaston and two referees are thanked for
28. J. H. Lawton et al., Nature 391, 72 (1998). comments.
29. Y. Haila and C. R. Margules, Ecography 19, 323
(1996). 7 October 1997; accepted 2 February 1998
Competition in Retinogeniculate Patterning
Driven by Spontaneous Activity
Anna A. Penn,* Patricio A. Riquelme, Marla B. Feller,†
Carla J. Shatz
When contacts are first forming in the developing nervous system, many neurons gen-
erate spontaneous activity that has been hypothesized to shape appropriately patterned
connections. In Mustela putorius furo, monocular intraocular blockade of spontaneous
retinal waves of action potentials by cholinergic agents altered the subsequent eye-
specific lamination pattern of the lateral geniculate nucleus (LGN). The projection from
the active retina was greatly expanded into territory normally belonging to the other eye,
and the projection from the inactive retina was substantially reduced. Thus, interocular
competition driven by endogenous retinal activity determines the pattern of eye-specific
connections from retina to LGN, demonstrating that spontaneous activity can produce
highly stereotyped patterns of connections before the onset of visual experience.
The circuitry of the adult nervous system of eye-specific layers is dependent on neu-
emerges from diffuse sets of early connections ral activity. When all action potentials in
(1). Although molecular cues guide initial the LGN are blocked by tetrodotoxin
axonal projections, activity-dependent com- (TTX), layers fail to form (13). To directly
petition is thought to sculpt precise connec- investigate the role of activity-dependent

iiii
7. , in (5), pp. 77–113.
8. and P. H. Williams, in (5), pp. 202–229.
9. J. M. Scott et al., Wildl. Monogr. 123, 41 (1993).
10. R. L. Pressey, C. J. Humphries, C. R. Margules, R. I.
Vane-Wright, P. H. Williams, Trends Ecol. Evol. 8,
124 (1993).
11. C. R. Margules, I. D. Cresswell, A. O. Nicholls, in
Systematics and Conservation Evaluation, P. L. Fo-
rey, C. J. Humphries, R. I. Vane-Wright, Eds. (Clar-
endon, Oxford, 1994), pp. 327–350.
12. A. O. Nicholls and C. R. Margules, Biol. Conserv. 64,
165 (1993).
13. D. P. Faith and P. A. Walker, Biodiv. Conserv. 5, 399
(1996).
14. J. R. Prendergast, R. M. Quinn, J. H. Lawton, B. C.
Eversham, D. W. Gibbons, Nature 365, 335 (1993).
15. A. T. Lombard, S. Afr. J. Zool. 130, 145 (1995).
16. A. P. Dobson, J. P. Rodriguez, W. M. Roberts, D. S.
Wilcove, Science 275, 550 (1997 ).
17. J. R. Prendergast and B. C. Eversham, Ecography
20, 210 (1997 ).
18. C. H. Scholtz and E. Holm, Insects of Southern Africa
(Butterworth, Durban, South Africa, 1995).
19. C. Muller, S. Freitag, C. H. Scholtz, A. S. van Jaars-
veld, Afr. Entomol. 5, 261 (1997 ).
20. K. J. Gaston, Oikos 61, 434 (1991).
21. International Council for Bird Preservation (ICBP), Put-
ting Biodiversity on the Map: Priority Areas for Global
Conservation (Birdlife International, Cambridge, UK,
1992); World Conservation Union (IUCN), Centres of
Plant Diversity: A Guide and Strategy for Their Con-
servation (IUCN, Richmond, VA, 1987 ).
22. A. G. Rebelo, Strelitzia 1, 231 (1994).
23. K. J. Gaston, Funct. Ecol. 6, 243 (1992).
24. A detailed list providing the degree of overlap be-
tween different taxa for complementary sets, rich-
ness hotspots, coldspots, and rare taxa is available
at www.sciencemag.org/feature/data/975464.sh/.
25. J. A. Harrison et al., Eds., The Atlas of Southern
African Birds (Birdlife South Africa, Johannesburg,
1997 ); R. M. Cowling and C. Hilton-Taylor, Strelitzia
1, 31 (1994); A. S. van Jaarsveld and S. L. Chown, S.
Afr. J. Sci. 92, 459 (1996).
tions (1–3). Activity-dependent competition
driven by visual input from each eye shapes
the ocular dominance columns of the visual
cortex (3). Likewise, activity-dependent
competition between motor neurons shapes
connections at the neuromuscular junction
(4). Yet precise connections can form in the
central nervous system before there is any
external sensory input (1, 5). Because many
of these connections are highly stereotyped,
it is generally thought that they are estab-
lished in response to molecular cues, rather
than by activity-dependent mechanisms (1).
However, endogenous neural activity could
pattern these connections through a compet-
itive mechanism.
In the mammalian visual system, retinal
ganglion cell inputs from each eye, initially
intermixed within the lateral geniculate nu-
cleus (LGN), become segregated during de-
velopment before vision (6–9). Correlated
bursts of spontaneous action potentials sweep
in “waves” across the retina and are transmit-
ted to the postsynaptic neurons in the LGN
during this segregation (10–12). Segregation
Howard Hughes Medical Institute and Department of
Molecular and Cell Biology, University of California,
Berkeley, CA 94720, USA.
* To whom correspondence should be addressed at 221
Life Sciences Addition, University of California, Berkeley,
CA 94720, USA. E-mail: apenn@uclink2.berkeley.edu
†Present address: National Institutes of Health, 36 Con-
vent Drive, MSC-4152, Bethesda, MD 20892, USA.
competition between the inputs from the
two eyes in the formation of eye-specific
LGN layers, we analyzed the effects of a
prolonged and selective blockade of the
retinal waves in ferret kits (Mustela puto-
rius furo).
Synaptic activation of neuronal nicotinic
receptors (nAChR) on ganglion cells is nec-
essary for the generation and propagation of
retinal waves (10). Fluorescence imaging ex-
periments (14) were used to monitor large
numbers of retinal ganglion cells simulta-
neously during bath application of cholin-
ergic agents to determine which compounds
would block the retinal waves and therefore
be potentially useful for in vivo intraocular
application. Periodic increases in intracellu-
lar calcium concentration {[Ca21]i}, which
are known to be associated with cholinergic
synaptic currents measured in ganglion cells
(10), are also correlated with bursts of action
potentials recorded from ganglion cells (Fig.
1A). Most action potential bursts occurred
simultaneously with changes in fluorescence
associated with waves propagating through
the surrounding tissue [96%; n 5 44 of 46
bursts recorded in nine cells from five retinas;
animals’ ages ranged from birth (P0) to post-
natal day 9 (P9)].
Both 10 mM nicotine and 100 mM curare
can block the periodic increases in [Ca21]i
(10), but these agents are short-acting (10,
15) and thus are poor candidates for intraoc-

2108 SCIENCE z VOL. 279 z 27 MARCH 1998 z www.sciencemag.org